Subject: Linux-Development Digest #784
From: Digestifier <Linux-Development-Request@senator-bedfellow.MIT.EDU>
To: Linux-Development@senator-bedfellow.MIT.EDU
Reply-To: Linux-Development@senator-bedfellow.MIT.EDU
Date:     Thu, 2 Jun 94 08:13:12 EDT

Linux-Development Digest #784, Volume #1          Thu, 2 Jun 94 08:13:12 EDT

Contents:
  Re: NFS hangs system 1.1.16 & 1.1.17 (Janne Sinkkonen)
  Re: SIGHUP - Where do we go from here? (Ken Pizzini)
  Device entry last use time (Dave Davey)
  CDU31A/33A driver with interrupt and DMA support uploaded (Corey Minyard)
  Re: XFree SVGA and 1.1.16 (Richard W. Altheide)
  Re: Linux 1.1.13 compiling-->DDIOCSDBG undefined in sbpcd.c (Eberhard Moenkeberg)
  Re: 1.1.16 NFS hangs on write when wsize >1k (Oliver Schoett)
  ptrace(...), and getting at registers (Michael Zastre)
  Re: Routing requests and 1.1.15 (njs@bcrvm1.vnet.ibm.com)
  Linux on a ThinkPad? (Wigs)
  Re: NFS hangs system 1.1.16 & 1.1.17 (Rene COUGNENC)
  1.1.16 kernel doesn't want to boot! (Eugene Tyurin)
  OI and UIB on linux with gcc-2.5.8 (Hima Sagar- The Humming Bird)
  Re: Writing a new file system, help (Andrew Davison)
  Re: SLIP problems with kernel 1.1.15 (Uri Blumenthal)
  Re: Why is my Emacs binary so big (2.2M) (David Dwayne Wiaizenegger)
  Re: How to print faster (Bram Bouwens)
  Busmouse on IRQ 2 works only with driver on IRQ 9 (Klaus Troja)
  net source.. (Jia-Ning Luo)
  Re: 1.1.16 kernel doesn't want to boot! (Rob Janssen)
  Re: Linux game development (Was Re: Why [DOS, W (Rob Janssen)
  Re: Frustrated with new kernels (Rob Janssen)

----------------------------------------------------------------------------

From: sinkkone@cc.Helsinki.FI (Janne Sinkkonen)
Subject: Re: NFS hangs system 1.1.16 & 1.1.17
Date: 1 Jun 1994 17:20:19 +0300

>in the office. When I try to rebuild the kernel on a mounted
>partition, it hangs both boxes - my suspicion is a memory deadlock
>of the fileserver box. NFS guru! Could this happen with the new
>8k sizes? Any way I can help track this down without dropping money
>for another screen or more memory?

I am not the guru, but have used 1.1.16 & 1.1.17 with 8k NFS. 8k NFS
write is broken in 1.1.16, and maybe also in 1.1.17, although the
latter has been working relatively well here (we have got only one
error message).

--
Janne



------------------------------

From: ken@coho.halcyon.com (Ken Pizzini)
Subject: Re: SIGHUP - Where do we go from here?
Date: 31 May 1994 21:26:33 GMT

In article <2setvn$fio@nwfocus.wa.com>, I <ken@halcyon.com> wrote:
> (Really gotta look at that source code...)

Okay, I finally went and did that -- it looks like a faithful
implementation of 1003.1-1988 in kernel/exit.c and
drivers/char/tty_io.c, except for the SIGCONT signals being sent
from do_tty_hangup() and disassociate_ctty() in
drivers/char/tty_io.c.  I can't see how these extra signals can
hurt anything, and I can see that these may be details that have
been added to the 199x revision of 1003.1, but I only have a
copy of the 1988 version, and the pedant in me feels the need to
point out this minor detail.

                --Ken Pizzini

------------------------------

From: daved@physiol.su.OZ.AU (Dave Davey)
Subject: Device entry last use time
Date: Wed, 1 Jun 1994 04:42:45 GMT

I have been trying to track down a parallel printer problem, and was
surprised to discover that "ls -lu /dev/lp*" is not giving last access
times on linux (1.1.15).  Is this an oversight or does it represent some
decision in principle?

Can I capture error messages from a PostScript printer on a parallel port?

------------------------------

From: minyard@crchh7b9 (Corey Minyard)
Subject: CDU31A/33A driver with interrupt and DMA support uploaded
Date: 1 Jun 1994 05:45:23 GMT
Reply-To: minyard@wf-rch.cirr.com

I have uploaded a driver for the CDU31A and CDU33A (it will auto-sense)
that has interrupt and DMA support added for boards that support it.
It is on sunsite.unc.edu in pub/Linux/Incoming and is named
cdu31a-v1.patch1.diff.  It is relative to v1.0 and 1.1 kernel releases.

Note that only DMA 3 is supported by default (and only on the Sony
interface boards), interrupts are turned off by default.  To turn them
on, you will have to edit the file and add the interrupt number to
the appropriate place in the table (it should be obvious where).
Interrupts are not turned on because the boards are shipped by default
with interrupts turned off.  If you are really brave, you can turn on
auto-interrupt detection.  My board doesn't support interrupts and it
detected my bus-mouse interrupt when I tried it, so I am dubious of it's
value.

Since the Pro Audio Spectrum supports neither DMA or interrupts, don't
enable any of these for that board.  It just won't work.  I don't think
the Soundblaster supports these, either, but I could be wrong on that
one.

If you try this. please mail me with the results, especially the speed
at which data transfers go.  The command:
        time dd if=/dev/cdu31a of=/dev/null count=10240 bs=2048
will transfer 20MB (20971520 bytes) of data and time the operation.

Thanks to Petteri Stenius, who helped a lot with the development of
the driver.

Corey
minyard@wf-rch.cirr.com

------------------------------

From: rwa@io (Richard W. Altheide)
Subject: Re: XFree SVGA and 1.1.16
Date: Tue, 31 May 1994 22:14:28 GMT
Reply-To: rwa@umr.edu

Richard W. Altheide (rwa@io) wrote:
: I have been running XFree 2.1.1 SVGA without difficulty using v1.1.11.  
: After moving up to v1.1.16 I find that the XVGA server hangs during 
: startup via xdm.  The screen goes blank like the X server is trying to 
   --- message text deleted ---

I have chased this to the serial mouse hanging.  I have got things 
working now most of the time by backing out the serial.c change of 
patch16 and using setserial 2.10 to initialize the serial ports.

The other system mentioned in the previous note that was not experiencing 
this problem has a ps2aux port type mouse rather than a serial mouse.


--
Richard W. Altheide, Sr. Systems Programmer   |   E-mail:          rwa@umr.edu
UM-Rolla Computing Services                   |   URL: http://www.umr.edu/~rwa
114 Mathematics/Computer Science Building     |   Telephone:    1.314.341.4841
Rolla, MO   65401-0249                        |   Fax:          1.314.341.4216

------------------------------

Date: Tue, 31 May 1994 01:21:10 +0200
From: Eberhard_Moenkeberg@p27.rollo.central.de (Eberhard Moenkeberg)
Subject: Re: Linux 1.1.13 compiling-->DDIOCSDBG undefined in sbpcd.c


Hello Elaine Walton and all others,

on 26.05.94 Elaine Walton wrote to All in USENET.COMP.OS.LINUX.DEVELOPMENT:

EW> I downloaded the Linux 1.1.13 kernel from tsx-11 last night and tried to
EW> compile it just now.  I get an error DDIOCSDBG not defined.  I searched all
EW> over the place and could not find it in my new or old code.  Can someone
EW> please tell me what its value should be?  Also what else might I be
EW> missing?

Two solutions (one is enough):

1. Install patch14.gz

2. Add the line
      #define DDIOCSDBG 0x9000
   somewhere at the beginning of sbpcd.c

BTW: If you get patch15.gz, do not compile it before you got patch16.gz. ;-)

Greetings ... Eberhard


------------------------------

From: Oliver Schoett <schoett@informatik.tu-muenchen.de>
Subject: Re: 1.1.16 NFS hangs on write when wsize >1k
Date: 1 Jun 1994 18:42:04 GMT

In article <1994May31.113456.10758@uk.ac.swan.pyr>
iiitac@uk.ac.swan.pyr (Alan Cox) writes:

 > Yep - 1.1.16 fixes some TCP bugs and in doing so broke fragmentation rather
 > subtly.. took me 4 hours to find. Fixes as usual on their way to Linus.

Thanks for the quick help.  1.1.17 works much better with 8k rsize and
wsize.

There is a problem, however, when reading from a large-ish (> 1MB)
NFS-mounted file and writing to disk at the same time: The commands

   tar xvzf NFS-file
   zcat NFS-file > local-file

both produce gzip crc errors sooner or later.  This happens with files
on both HP and Sun servers.  Commands which simply read the files
(e.g., sum, tar tvzf, zcat >/dev/null) do not exhibit any problem.
The problem does not occur with rsize = wsize = 1K.


Oliver Schoett     Inst. f. Informatik (H2), Technische Univ. Muenchen
                   Postfach 20 24 20,     8000 Muenchen 2,     Germany
schoett@informatik.tu-muenchen.de    phone +49 89 2105-8474  fax -8207

------------------------------

From: zastre@sanjuan.UVic.CA (Michael Zastre)
Subject: ptrace(...), and getting at registers
Date: Wed, 1 Jun 94 17:10:55 GMT


Has anyone out there had much experience with "ptrace(...)" on Linux?
I'm running Slackware 1.2.0, and was overjoyed to find a man page on
the function.  However, it sheds no light onto how to look at the register
contents.  The include file "linux/ptrace.h" helps a bit, but it uses
a macro in a comment that doesn't seem to exist anywhere (PT_EXCL).

Mike Zastre
zastre@sanjuan.uvic.ca


------------------------------

From: njs@bcrvm1.vnet.ibm.com
Subject: Re: Routing requests and 1.1.15
Date: Wed, 1 Jun 1994 19:15:40 GMT
Reply-To: njs@watson.ibm.com

In <1994May28.204721.1073@troehl.hanse.de>, thomas@troehl.hanse.de (Thomas Roehl) writes:
>Mike Dowling (mike@MooCow.math.nat.tu-bs.de) wrote:
>: When I boot I get 3 messages to the effect
>
>: Warning: obsolete routing request made.
>
>: It looks to me as if this results from the following 3 lines in my rc.inet1
>
>: /sbin/route add localhost
>: /sbin/route add tubs-net eth0
>: /sbin/route add default gw 134.169.2.1 1
>
>: If so, can anyone please tell me how this is now supposed to be done?
>: (Everything nevertheless works fine.)
>
>...then I recompiled route.c (dont't remember 
>exactly from where I got it) and no more messages.

route.c automatically adapts based on a defined symbol to
putting route requests through using a new ioctl.  Fix this
by recompiling route.c.
--
Nick Simicich - njs@bcrvm1.vnet.ibm.com - njs@watson.ibm.com
                - njs%scifi.uucp@uunet.uu.net

------------------------------

From: wiegley@phakt.usc.edu (Wigs)
Subject: Linux on a ThinkPad?
Date: 1 Jun 1994 13:36:05 -0700


I know this has been asked before... but I can't find the answer in the
compatibility list.

Does anyone have Linux running on an IBM ThinkPad?  If you do then what
problems did you experience? I saw a lot of posts by people having drive
problems, have these been fixed by now or are their solutions for these
problems if they still exist?

Thanks,

- Jeff Wiegley
wiegley@usc.edu

------------------------------

From: rene@renux.frmug.fr.net (Rene COUGNENC)
Subject: Re: NFS hangs system 1.1.16 & 1.1.17
Date: 1 Jun 1994 18:37:32 GMT
Reply-To: cougnenc@hsc.fr.net (Rene COUGNENC)

Ce brave Alan D. Peckham ecrit:


> This box is presently a small headless 386/40 4MB memory and
                                        ^^^^

Read the thread in this group, "kernel 1.1.14 (&13) + xiafs segfaults rpc.nfsd",
(xiafs is not the problem in fact)... There is indeed a problem with
nfsd, the new kernels and 386 processors :-(

--
 linux linux linux linux -[ cougnenc@renux.frmug.fr.net ]- linux linux linux 

------------------------------

From: gene@insti.physics.sunysb.edu (Eugene Tyurin)
Subject: 1.1.16 kernel doesn't want to boot!
Date: 01 Jun 1994 18:02:54 GMT


        I posted this question some time ago to comp.os.linux.help,
but since nobody helped me, I decided to post here.

        I tried to upgrade from 1.1.12 to 1.1.16. Kernel got compiled
and installed with absolutely no problems. But when I rebooted
machine, it loaded compressed kernel and gave me

                crc error
                System halted

upon uncompressing. I  repeated the whole procedure (configure,
compile, install) again, but (alas!) with the same result. What's wrong?

--
--                       MIME mail is welcome                      --
Active Ingredient: Eugene Tyurin  <gene@insti.physics.sunysb.edu>
                   http://insti.physics.sunysb.edu:80/~gene/plan.html
                   Bio-degradable, made of 100% recycled substance.
                   Handle with care: fragile, toxic and flammable.



------------------------------

From: vsvankay@acs.ucalgary.ca (Hima Sagar- The Humming Bird)
Subject: OI and UIB on linux with gcc-2.5.8
Date: Wed, 1 Jun 1994 23:20:24 GMT

Hi,

I have been trying to use OI and UIB (tm ?)  from ParcPlace on a linux
box. The documentation says that the version is fully functional, yet
I am unable to use it.

Description : 
I can fire up uib, and build an interface, such as the
filepreviewer example from the documentation. I can save it too. It
generates various files, and the makefile.

But when I compile, I get errors with the typedef of member functions
in defs.H file. 

Compilation log :
==========================================================
g++ -c -g  -DIS_MAIN_APP    previewer.C
In file included from /usr/include/OI/d_tech.H:28,
                 from /usr/include/OI/subclass.H:28,
                 from /usr/include/OI/oi.H:32,
                 from previewer.C:23:
/usr/include/OI/defs.H:515: typedef name may not be class-qualified
/usr/include/OI/defs.H:522: typedef name may not be class-qualified
/usr/include/OI/defs.H:529: typedef name may not be class-qualified
/usr/include/OI/defs.H:536: typedef name may not be class-qualified
/usr/include/OI/defs.H:543: typedef name may not be class-qualified
/usr/include/OI/defs.H:550: typedef name may not be class-qualified
/usr/include/OI/defs.H:557: typedef name may not be class-qualified
/usr/include/OI/defs.H:565: typedef name may not be class-qualified
/usr/include/OI/defs.H:576: typedef name may not be class-qualified
/usr/include/OI/defs.H:586: typedef name may not be class-qualified
/usr/include/OI/defs.H:593: typedef name may not be class-qualified
/usr/include/OI/defs.H:600: typedef name may not be class-qualified
/usr/include/OI/defs.H:607: typedef name may not be class-qualified
/usr/include/OI/defs.H:614: typedef name may not be class-qualified
/usr/include/OI/defs.H:621: typedef name may not be class-qualified


.. etc etc..
===================================================================

I tried various combinations of -D_AIX, -Dultrix
-Dlinux etc with the CCFLAGS, but I am still unable to compile.

I am using gcc-2.5.8.

p.s. I tried to compile the already existing files in the demos
directory of uib. They too give the same compilation error.

Any help will be greatly appreciated. 

Thanx
vidya



-- 
Vidya 'Hima' Sagar  -- The Humming Bird
vsvankay@acs.ucalgary.ca (or) vankay@enel.ucalgary.ca

------------------------------

From: davison@bruce.cs.monash.edu.au (Andrew Davison)
Subject: Re: Writing a new file system, help
Date: 1 Jun 1994 08:10:10 GMT

Ben Frank (u9219765@sys.uea.ac.uk) wrote:
: Hi,
:    I'm trying to write a new file system (Third year porject) I seem to
: remember that someone was writing a "HowTo write a file system" document,
: I can't find it anywhere, nor any reference to it. Does it exist, if not is
: there anything that would be useful to look at ? I have read/am reading all
: the stuff from the LDP
: Cheers.

:       Ben Frank.

You'll probably want to take a look at the userfs package on sunsite before
getting to stuck into it. That will let you iron out most of your problems
before you start running your machine through copious power cycles.

Andy

------------------------------

From: uri@watson.ibm.com (Uri Blumenthal)
Crossposted-To: comp.os.linux.help
Subject: Re: SLIP problems with kernel 1.1.15
Date: 1 Jun 1994 22:16:08 GMT

Get a newer DIP - "dip337b-uri.tgz" is verified to work with
1.1.16 kernel for both dial-in and dial-out. Behold the
change in /etc/diphosts format. Enjoy.

------------------------------

From: Dave.Waizenegger@launchpad.unc.edu (David Dwayne Wiaizenegger)
Crossposted-To: comp.os.linux.help
Subject: Re: Why is my Emacs binary so big (2.2M)
Date: 2 Jun 1994 02:13:57 GMT


Well, I didn't read the original post, but it seems to be about size
optimization.  Some suggestions for decreasing the size of your binary: 
1) Use the optmize flag (-O or -O2) depending on the optimization you want.
2) Don't use machine optimization that are likely to increase size (-m486)
3) Use shared libs.  Make sure you don't force or imply static libs.
4) Strip debugging info at llink time (-s flag on ld) or use the 'strip'
command on the final binary.  This can make a big difference.

Of course there are lots of things, but these things are some important ones.
Hope this helps someone.   Regards,
                                Dave

--
==============================================================================
 \ The above does not represent OIT, UNC-CH, laUNChpad, or its other users. /
   ------------------------------------------------------------------------

------------------------------

From: d70@nikhef.nl (Bram Bouwens)
Subject: Re: How to print faster
Date: 2 Jun 94 09:29:13 GMT
Reply-To: d70@nikhefh.nikhef.nl (Bram Bouwens)

In article <nigel.770473352@gamble> nigel@gate.net (Nigel Gamble) writes:
=In <CqII9r.47F@pe1chl.ampr.org> rob@pe1chl.ampr.org (Rob Janssen) writes:
=
=>Interesting...  apparently the printer delays quite a while before it
=>releases busy after sending the ACK.
=>The PC parallel port interrupt design is braindead, in that it sends an
=>interrupt before the device is ready for more data.
=
=It's up to the printer to indicate when it is ready for the next
=data byte to be strobed in by asserting ACK and dropping BUSY.  It is
=the ACK that causes the interrupt, but the fact that this comes before
=the printer drops busy is not supposed to make any difference - by
=asserting ACK, the printer has declared itself ready to receive the
=next character.  It's the particular printer design that's brain-dead
=if it's not behaving like this.

As far as I know ACK simply means that the printer has received the
character, so the datalines no longer have to be stable. If the
8255 were used in its strobed output mode, it would even wait for
this ACK to release the STB, which is a good thing with long cables.
Some printer manuals do specify that ACK comes soon after STB, but
not how much time elapses before BUSY is dropped, as this depends
on what the printer is actually doing.
Simply put: ACK is not suitable to trigger interrupt driven output.

Bram
======================================================================
  Bram Bouwens <Bram.Bouwens@nikhef.nl>
  NIKHEF-H, National Institute for Nuclear and High-Energy Physics
======================================================================

------------------------------

From: klaus@mail.sz.etc.tu-bs.de (Klaus Troja)
Subject: Busmouse on IRQ 2 works only with driver on IRQ 9
Date: Thu, 2 Jun 1994 07:41:29 GMT

Hi there!

I have got an ATI Bus Mous on a Graphics Vantage. The driver sources for bus 
mice are set for IRQ 5. I changed them to IRQ 2 without success.
You have to set it to IRQ 9 in the driver to get the mouse working. Because of 
the cascaded two Interrupt Controllers IRQ 2 on the first refers to IRQ 9 on 
the second controller which is asked first.

Klaus

------------------------------

From: deer@tem.nctu.edu.tw (Jia-Ning Luo)
Subject: net source..
Date: 2 Jun 1994 11:08:09 GMT


  Where to get the newest sources/include file /library of the
 
  net 3.x  ?   :-)

----

                           Jia-Ning Luo,   deer@ios706.ncu.edu.tw

------------------------------

From: rob@pe1chl.ampr.org (Rob Janssen)
Subject: Re: 1.1.16 kernel doesn't want to boot!
Reply-To: pe1chl@rabo.nl
Date: Thu, 2 Jun 1994 08:10:26 GMT

In <GENE.94Jun1140254@brook.physics.sunysb.edu> gene@insti.physics.sunysb.edu (Eugene Tyurin) writes:


>       I posted this question some time ago to comp.os.linux.help,
>but since nobody helped me, I decided to post here.

>       I tried to upgrade from 1.1.12 to 1.1.16. Kernel got compiled
>and installed with absolutely no problems. But when I rebooted
>machine, it loaded compressed kernel and gave me

>               crc error
>               System halted

>upon uncompressing. I  repeated the whole procedure (configure,
>compile, install) again, but (alas!) with the same result. What's wrong?

Probably you did not run LILO, or did not run it properly.  See
the LILO documentation.

Rob
-- 
=========================================================================
| Rob Janssen                | AMPRnet:   rob@pe1chl.ampr.org           |
| e-mail: pe1chl@rabo.nl     | AX.25 BBS: PE1CHL@PI8UTR.#UTR.NLD.EU     |
=========================================================================

------------------------------

From: rob@pe1chl.ampr.org (Rob Janssen)
Subject: Re: Linux game development (Was Re: Why [DOS, W
Reply-To: pe1chl@rabo.nl
Date: Thu, 2 Jun 1994 08:14:31 GMT

In <2sj9hi$ltm@engnews2.Eng.Sun.COM> sullivan@thresher.Eng.Sun.COM (John J. Sullivan) writes:

>The main objection that I see to using Linux is that games developed 
>under it cannot be run directly from DOS, and generally require Linux
>to already be installed on a user's machine.  Here's what I suggest:

>1. Develop a version of LILO that runs as a DOS executable and can boot
>   a Linux kernal image from a DOS file.
>2. Develop a minimal Linux kernel that, when booted, mounts the DOS disk
>   it's root directory, and runs the game executable.

Look at the "umsdos" filesystem.  It supports having a Linux system on
a DOS disk.  Programs to boot Linux from DOS exist as well.

Rob
-- 
=========================================================================
| Rob Janssen                | AMPRnet:   rob@pe1chl.ampr.org           |
| e-mail: pe1chl@rabo.nl     | AX.25 BBS: PE1CHL@PI8UTR.#UTR.NLD.EU     |
=========================================================================

------------------------------

From: rob@pe1chl.ampr.org (Rob Janssen)
Subject: Re: Frustrated with new kernels
Reply-To: pe1chl@rabo.nl
Date: Thu, 2 Jun 1994 08:15:57 GMT

In <2sje4b$ntu@charm.magnus.acs.ohio-state.edu> ewalton@magnus.acs.ohio-state.edu (Elaine Walton) writes:

>I have tried to compile the new kernels from off of tsx-11--no luck.
>I am really confused how others are able to compile, yet I am not.
>Maybe it's because I'm trying to go from 1.0.0 directly to 1.1.13-6?
>Either way, I have pretty much given up..

Many people also have pretty much given up replying to vague complaints
like that.

Rob
-- 
=========================================================================
| Rob Janssen                | AMPRnet:   rob@pe1chl.ampr.org           |
| e-mail: pe1chl@rabo.nl     | AX.25 BBS: PE1CHL@PI8UTR.#UTR.NLD.EU     |
=========================================================================

------------------------------


** FOR YOUR REFERENCE **

The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:

    Internet: Linux-Development-Request@NEWS-DIGESTS.MIT.EDU

You can send mail to the entire list (and comp.os.linux.development) via:

    Internet: Linux-Development@NEWS-DIGESTS.MIT.EDU

Linux may be obtained via one of these FTP sites:
    nic.funet.fi				pub/OS/Linux
    tsx-11.mit.edu				pub/linux
    sunsite.unc.edu				pub/Linux

End of Linux-Development Digest
******************************
